Lumatik Film & Animation

🎅 Santa vs. DATA-SANTA! 🎄✨ Auch dieses Jahr hatten wir die große Freude, für NetCologne einen weihnachtlichen Animationsfilm zu produzieren – mit einer ordentlichen Prise Kölner Lokalkolorit und Humor! 😄 Dieses Mal dreht sich alles um das neue ökologische Datacenter, das wir in einer kreativen Story eingebettet haben. Natürlich spielt die Geschichte wieder im beliebten Heinzelverse – unserer etablierten NetCologne-Storywelt. Und das Beste? Wie immer kommt die Erzählung in Reimform daher. 🎥🎶 Es ist einfach jedes Mal ein Geschenk, mit einem Kunden zusammenzuarbeiten, der uns so viel Vertrauen in unsere kreative Arbeit schenkt. 💡❤️ #Weihnachten #Animation #Köln #brandedentertainment #contentmarketing #storytelling

Generative AI is cool and all, but procedural 3D modeling just hits different. Check out this Houdini setup by Pepe Buendia. Why this is cool: Instead of manually placing every building and car, this system generates an NYC-style city that builds itself -- automatically spawning buildings with unique variations, plus a traffic system where cars actually obey traffic lights and avoid crashing into each other (all orchestrated through houdini VEX code). Each car knows exactly which road it's on and makes real-time decisions about navigation. The whole thing runs on custom algorithms that handle everything from road directions to building placement. If you did this with generative AI, even approaches to conditioning and taming the chaos -- you'd still get a ton of variance which might be an issue for visual and spatial consistency. To me this is procedural modeling at its finest -- precise control meets infinite scalability. Immensely useful for media & entertainment, but also synthetic training data for robotics and augmented reality. I of course can't wait until multimodal LLMs to have the spatial understanding to write these kind of controllable 3D worlds for us. Claude counting 2D pixels on a screen is a step in this direction. Soon they'll be able to manipulate 2D ortho viewports, and eventually spawn procedural 3D worlds.

It is amazing to see that the trend of NPR (non-photorealistic rendering) continues to go strong. It adds so much more creative choices to the art of 3D design and animation.

Es ist alles gesagt und nun Zeit zu handeln.

It’s been barely a year since Wonderstudio was the talk of the town for its ability to motion-track a person in a video and automatically composite a 3D character in their place—one that still needed to be modeled and rigged beforehand. And now, this: A single picture is all you need?

For years we have been using the software PreProducer Software UG for preparing, organizing and budgeting our film projects. It was especially useful that it integrated so well with our favorite script writing software Dramaqueen. Now Preproducer has integrated data exchange with the LockitNetwork App. A great way to update your production data with info from the set via the script supervisor. Extremely helpful also if relevant data needs to get to the VFX team feast. Great move, Preproducer! #filmproduction #scriptsupervision #vfx https://lnkd.in/gtYP5An4

Today a new AI company, backed by major players in the AI Game came out of stealth and presented their first generative AI Model Flux.1. It's not just another picture generation model. It gave all the established players like Midjourney, Dall•E and Stable Diffusion a run for their money, right out of the gate. And a video generation Model is already announced. What I find exiting beyond that, is the fact that it comes out of Germany. It's called Black Forest Labs for a reason - it's based in Freiburg. To celebrate this, here is a comparison of how well the most popular image generators and the new kid on the block do when asked to produce a typical southern-German "Brotzeit" platter. #generativeAI #AI Black Forest Labs , Stability AI , Midjourney , OpenAI

#SIGGRAPH2024, and one of the greatest advancements in #GenAI technology that can no longer be ignored, high-end #3Dcontent creation. *NOTE: Updates in comments, from #Shutterstock unveiling its own #TextTo3D foundational model, generated from its catalog of #licensedcontent. Creating complex 3D worlds and objects can be tough, requiring lots of skills and effort with the current digital tools. To make this easier, they have developed CLAY—a powerful tool designed to transform your ideas into detailed 3D models effortlessly. CLAY: Controllable Large-Scale Generative Model for Creating High-Quality 3D Assets Easy Input Methods: - You can start with simple text descriptions or images. - Use basic 3D shapes and forms like multi- view images, #voxels (3D pixels), bounding boxes, #point clouds, and more. Advanced Technology: - CLAY uses a sophisticated AI model called a Variational Autoencoder (VAE) paired with a Diffusion Transformer (DiT). These help CLAY understand and generate 3D shapes by learning from a huge variety of 3D models. - The tool also uses #neuralradiancefields (#NeRF) to create smooth, complete surfaces, and it leverages advanced transformer blocks to work in a simplified, abstract space. Massive Training Data: - They trained CLAY using a gigantic dataset of 3D models, ensuring it understands a wide range of shapes and textures. This dataset was processed meticulously to provide the best learning material for the AI. *BETTER DATA = BETTER RESULTS* High-Quality Textures: CLAY can create highly detailed and realistic textures for your 3D models. These textures support physically-based rendering (#PBR), which helps make the final result approach something lifelike (professional finesse is still a requirement). You get #textures with different properties like diffuse, roughness, and metallic finishes, all in high resolution (2K). With CLAY, digital businesses can streamline the creation process, reduce reliance on highly specialized skills, and foster an environment where creativity can flourish without technical barriers. At the same time, there's no question that digital businesses like #videogame and #vfx studios, will leverage and/or expand upon these techniques, further scaling the approach from original IP/assets, to greatly enhance the quality of such a foundational model type. Again, #AI won't take your job, but someone using AI absolutely will. The differentiating factor here, is how a human can creatively apply the technology to clever business strategies adapted for the bleeding-edge of present domains, such as #advertising, #branding, #marketing, #metaverse, #ecommerce, etc. TRY IT on HF: https://lnkd.in/g95yzP_D Project page: https://lnkd.in/g3q3Q-VM #Github: https://lnkd.in/gF6BZJHb

Exciting news from Stability AI! They have just published a groundbreaking paper on their innovative "4D" video generation model. This novel approach transforms any 2D footage into a 3D scene, enabling re-shooting from different angles with a virtual camera. While the technology is still in its early stages and there is room for improvement in achieving high-quality results, it undoubtedly represents a promising avenue for future exploration. https://lnkd.in/dbH77YvV #ai #film #3d